INDEX

Explanations

mentions of gender identity and sexual orientation, particularly related to women and the LGBTQ+ community

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 23-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L23R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.23.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

jom

-0.15

 Habit

-0.15

pcf

-0.14

åıİ

-0.14

 homosexuals

-0.14

 Habitat

-0.14

ibu

-0.13

zioni

-0.13

-gradient

-0.13

ê¹¨

-0.13

POSITIVE LOGITS

 identifying

0.30

 identification

0.28

 identify

0.28

 identity

0.28

 identities

0.26

 Identification

0.26

 ident

0.26

 identifies

0.26

identity

0.25

 Identify

0.25

Activations Density 0.117%

mentions of gender identity and sexual orientation, particularly related to women and the LGBTQ+ community

No Comments

No Known Activations